Instructions to use zai-org/GLM-5.2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use zai-org/GLM-5.2 with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="zai-org/GLM-5.2")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForMultimodalLM

tokenizer = AutoTokenizer.from_pretrained("zai-org/GLM-5.2")
model = AutoModelForMultimodalLM.from_pretrained("zai-org/GLM-5.2")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Inference
HuggingChat
Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use zai-org/GLM-5.2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "zai-org/GLM-5.2"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "zai-org/GLM-5.2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/zai-org/GLM-5.2

SGLang

How to use zai-org/GLM-5.2 with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "zai-org/GLM-5.2" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "zai-org/GLM-5.2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "zai-org/GLM-5.2" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "zai-org/GLM-5.2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use zai-org/GLM-5.2 with Docker Model Runner:
```
docker model run hf.co/zai-org/GLM-5.2
```

We need some Air or at least some Flash

by jacek2024 - opened about 22 hours ago

Discussion

jacek2024

about 22 hours ago

GLM-4.5-Air and GLM-4.7-Flash are among my favorite models, please consider bringing back models for local users.

AXYZE9

about 22 hours ago

GLM-5.2 Air would be amazing for subagents and local use!

v3ss0n

about 21 hours ago

Please DGX Spark user here, we got shamed coz we don't have enough models to run. Air Please..

DaveTheKiller

about 11 hours ago

An updated GLM-4.7-Flash would be amazing! GLM-5.2 Air?!

I'll take what I can get though... I appreciate all open source Z.ai models (thank you!)

soyalemujica

about 5 hours ago

What about a 27B dense one ? I mean, Qwen 3.6 27B is the king here, I am certain zai-org can give us a 27B dense model for agentic coding and beat Qwen !

cibernox

about 5 hours ago

•

edited about 5 hours ago

I also think that if this excellend model could be distilled and concentrated into 30B and 60B models (for us peasants with, respectvely, 1x24gb cards and 2x24gb cards), that would be idea.
There seems to be a big gap between 30B models that can run on a single gaming GPU and 120+B models that requires a $15k system with an RTX6000 that nobody is targeting. Something around 60B.

mostfake

about 3 hours ago

A dense 30b model could provide a good distill of this. Not sure what your company goals are but I would love it from a local perspective.

raulgdm

about 3 hours ago

+1 for a local 27 or 30b ai beast from you Guys!. Would be much apreciated for the gpu poor .

void-mckenzie

about 3 hours ago

We'd love a Flash or an Air model please!! Thank you for the models

rmhubbert

about 2 hours ago

Another +1 for a new Flash or Air. 120B MOE, or 80B dense is the sweet spot for myself, but anything that can be run locally would be amazing.

juanquivilla

about 1 hour ago

Another huge +1 here for a smaller Air/Flash model of GLM. This would be incredible and a great stepping stone to the big model.

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment